Policymakers persist in attempts to promote transit as a means of reducing automobile use and its attendant problems, including air pollution. In general, such attempts have been unproductive (Pickrell, 1992; Kain, 1991; Gomez-Ibanez, 1985; Richmond, 1998). The long-term trend in transit ridership has been flat and has declined significantly since 1990, despite decades of huge Federal and State subsidies. Arguments that eliminating subsidies to automobiles would change the modal split more in favor of transit are unconvincing because the price elasticity of demand for auto use remains low (Ingram and Liu, 1999), even in the long run. Where transit ridership has remained relatively stable, the explanation has been a demand from people without cars (e.g. working immigrant women in one-car households) rather than trip diversion from automobiles. Paradoxically, transit ridership has dropped even in the new rail cities; the reason is that the huge costs of building and expanding rail lines have cannibalized the bus system so that the modest expansion in rail ridership has been swamped by a more dramatic decline in bus ridership as services are cut. The costs of rail systems (capital costs often several times initial projections and high operating and maintenance costs) are difficult to justify: "Oft-quoted gross ridership figures for light rail may seem impressive. Total systems perspectives are however needed to make us realize that the total impact on public transit ridership is not only slight but that equal or better results can be obtained from relatively minor adjustments of fare levels and low cost improvements to existing bus services" (Richmond, 1998, p. 95).